Veto Players in Post-Conflict DDR Programs: Evidence From Nepal and the DRC

ABSTRACT Under what conditions are Disarmament, Demobilization and Reintegration (DDR) programs successfully implemented following intrastate conflict? Previous research is dominated by under-theorized case studies that lack the ability to detect the precise factors and mechanisms that lead to successful DDR. In this article, we draw on game theory and ask how the number of veto players, their policy distance, and their internal cohesion impact DDR implementation. Using empirical evidence from Nepal and the Democratic Republic of Congo, we show that the number of veto players, rather than their distance and cohesion, explains the (lack of) implementation of DDR.


Introduction
Disarmament, Demobilization and Reintegration (DDR) programs have become essential components of post-conflict peacebuilding. By collecting and disposing of weapons and ammunition from combatants, discharging ex-combatants from their respective armed groups and reintegrating them into civilian communities, DDR programs are widely regarded a core necessity to foster peace after war (Knight and Özerdem 2004;Muggah and O'Donnell 2015).
Increasingly, scholarly research studies the causes and consequences of successful DDR programs. Four recurring themes dominate the debate. First, studies spell out a number of technical requirements for successful DDR, such as arms reduction (e.g. Colletta, Kostner, and Wiederhofer 1996;Muggah 2005), sequencing, or creating economic opportunities for ex-combatants through education projects or small business loans (Banholzer 2014;Munive and Stepputat 2015). Second, studies discuss the vital role of international actors for DDR success, such as of the United Nations (UN) or international development and humanitarian organizations (Ball 1997). Third, studies examine the relationship between successful DDR and Security Sector Reform or SSR (Ball 1997;Mobekk 2009), noting that DDR is often an essential prior step to SSR that helps to increase trust between warring parties and creates a culture in which the use of weapons becomes unlikely (Muggah 2005). Finally, studies increasingly focus on the role of gender identities (Colletta, Kostner, and Wiederhofer 1996;Kaufman and Williams 2015), emphasizing that DDR programs also fail as they ignore the special needs of female combatants (Jennings 2009).
But despite the increasing attention in the literature, what works in post-conflict DDR and what doesn't is often not very clear. This lack of systematic knowledge on what factors determine the variation in DDR success is also due to the fact that past research is dominated by under-theorized single case studies or policy-oriented best-practice guidelines. While studies offer fine-grained insights into the dynamics of DDR in a few prominent cases, they are little integrated into the wider theoretical literature on post-conflict peacebuilding and lack the ability to detect the precise variables and causal mechanisms affecting the success of DDR (cf. Ansorg, Haass, and Strasheim 2013).
To tackle this shortcoming, we examine the utility of veto player theory in identifying the conditions that increase the likelihood of successful DDR implementation. Veto player theory is part of the wider field on game theory. It rests on a number of assumptions, which include the following: veto players take rational decisions to increase their benefit; they are aware of the rules of the game and they have certain preferences over outcomes; veto players adopt one or more strategies to tackle the problem; outcomes are dependent on the preferences of actors and institutions of a political system (Tsebelis 2002, 17ff.). Veto player theory can be adopted for a number of situations, including policy decisions on DDR after large-scale conflict.
In this paper, we explore how (1) the number of veto players, (2) their distance, captured through the aspect of ethnic cleavages, and (3) their internal cohesion, captured through the degree of fractionalization and leadership continuity, affect DDR implementation. We thereby follow past research and consider such implementation to be successful if most or all of the anticipated number of weapons and ammunitions held by combatants are turned in and/or destroyed; most or all of the expected number of ex-combatants have participated in a demobilization program; and most or all of the expected number of ex-combatants are reintegrated into society (cf. Joshi, Quinn, and Regan 2015).
We use the cases of DDR in Nepal and the Democratic Republic of Congo (DRC) to illustrate the mechanisms predicted by veto-player theory. Both cases offer fruitful empirical insights into key components of the veto player approach: Nepal saw a select number of veto players with similar policy goals and a moderate degree of fractionalization; while the DDR process in the DRC had to deal with a large number of distanced and fractionalized players. As selecting cases from distinct world regions 'still constitutes the exception to the rule' when analyzing policies in the Global South (Basedau and Köllner 2007, 112), we expect that our results do not reflect factors that are specific to only one region, but represent a more generic explanation of success and failure of the implementation of DDR. Having said that, we explicitly stress that the different context conditions in the two cases do not allow us to employ an explicitly comparative methodological strategy. These distinct context conditions include, but are not limited to, the regional dynamics in the case of the DRC as opposed to the conflict in Nepal that was confined in its national borders; or the role natural resources played in the DRC without taking on a similar significance in Nepal. Instead of explicitly comparing the two cases, we rather use them to illustrate and highlight theoretical mechanisms derived from the veto player approach. The two cases provide a useful empirical background to test our theoretical assumptions and to reveal limitations of the theory, in particular in a highly dynamic and volatile political environment.
For the case study on Nepal's DDR program, our analysis chiefly rests on new information gathered through semi-structured interviews and focus group discussions. In total, 57 interviews and three focus group discussions were conducted during two rounds of field work in the fall of 2015 and spring of 2017. Interviewees included current and previous members of government; high-ranking officers in the Nepal Army (NA); former child soldiers, (female) combatants, and commanders of the Communist Party of Nepal (Maoist) or CPN (M) rebels; as well as members of civil society and the international community. Focus groups were held with former combatants integrated into the NA as well as 'regular' NA soldiers. Most individuals spoke under the condition of anonymity, and about half of the discussions took place with the help of a translator. The case study additionally draws on existing research, policy reports, and news articles.
For the case study on the DRC, we had to rely on primary and secondary sources, as field work originally planned for March 2017 was too dangerous to conduct, due to ongoing tensions following the refusal of President Kabila to step down after his constitutionally mandated two-term limit expired in December 2016. We thus rely on programme reports and evaluations by the World Bank and other actors, independent assessments by third actors, as well as single case studies and policy reports. We are aware that the information gathered through these sources is limited in comparison to fieldwork-generated data. However, as the Congolese DDR process is well assessed by donors and actors, we believe that this provides us with the necessary information to conduct our analysis.
The remainder of this article is structured as follows. In section two, we elaborate on the theoretical assumptions of veto player theory in DDR programs. In section three, we briefly introduce the background to each case and DDR program, before providing empirical evidence on the causal mechanisms linking our independent and dependent variables. In sum, our findings suggest that while the number of veto players is a sound explanation for the (lack of) implementation of DDR in both cases, our results are more ambiguous when it comes to the policy distance and cohesion of players. Section four concludes by outlining avenues for future research.

Veto players and the implementation of DDR programs
What explains the successful implementation of DDR? Tsebelis (e.g. 1995Tsebelis (e.g. , 2002 offers a useful framework to study this puzzle by analyzing how different configurations of veto players affect the possibility of policy change. He defines veto players as 'individual or collective actors whose agreement is necessary for a change of the status quo' (Tsebelis 2002, 19), and argues that every political system has a certain configuration of players that makes the departure from the status quo ('policy change') more or less likely. Here, we define the successful implementation of DDR as such policy change or as movement away from the status quo that is the presence of fully armed and mobilized warring parties. Veto players are thus those actors who have the skills and ability to successfully veto the implementation of DDR. Following Tsebelis, we can formulate three hypotheses about what configurations of players should affect the successful implementation of DDR.
First, we expect a higher number of veto players to decrease the possibility of policy change: If decision-making over policy change is delegated to one player, the status quo follows her policy positions, but as the number of players required to agree to a movement away from the status quo increases, the ability for policy change decreases. Many actors can be veto players in conflict-affected societies, and Cunningham (2013, 38) defines veto players in such contexts as groups 'that have the ability to block an end to war' even if other actors agree to implementing a peace deal. Such players include governments and rebel groups, as well as external states supporting these parties. Cunningham (2013, 41) notes that not all rebel groups are automatically veto players, and that those that have more troops or better military equipment and training are more likely to be veto players. Second, following the veto player approach we expect that an increase of the policy distance of players required to change the status quo reduces the likelihood of policy change (Tsebelis 1995, 298): More diverse preferences of players shrink the size of the bargaining range and the chance of coming to terms on a deal all players find acceptable (Cunningham 2006, 881). While veto player approaches in the political science literature often use ideological preferences of players on a left-right continuum to conceptualize the polarization of players, Cunningham (2006Cunningham ( , 2013 suggests to use a different measurement to capture polarization in conflict-affected societies. He argues that splinter rebel groups are unlikely to bring a diverse set of policy preferences to the table, meaning their position should be more congruent with other players in the game. Further ways of conceptualizing polarization suggested in the peace and conflict literature are ethnic cleavages (Svensson 2009) and indivisible stakes in war (Cunningham 2011, 90).
H2: The likelihood of successfully implementing post-conflict DDR programs decreases as the policy distance between veto players increases. Third, following veto player theory we expect that the internal cohesion of collective veto players, defined as the 'similarity of policy position' of the constituent units of players (Tsebelis 1995, 301), affects policy change. According to Tsebelis, less cohesive players increase the possibility for incremental change as small deviations from the status quo may be approved by the majority. We expect that this assumption does not translate to highly volatile post-conflict situations. In fact, Cunningham (2006) proposes capturing player cohesiveness in such contexts by studying whether or not a party avoided fractionalization and leadership change during war. Following this strategy, we expect that if the claims of a cohesive party are met in peace negotiations, it will likely stop fighting and adhere to the terms of DDR negotiated in a peace deal. A centralized authority and lack of fractionalization can encourage a better implementation as parties execute orders along the lines of authority. Fractionalized parties might, however, continue fighting if one fraction feels under-represented or thinks it can get a better deal in future. Fractionalized parties may thus be less likely to implement change.
H3: The likelihood of successfully implementing post-conflict DDR programs decreases as the cohesion within collective veto players decreases.

Veto players and post-DDR in Nepal
Nepal's 'People's War' broke out in February 1996, when Maoist insurgents began attacking police stations and government offices particularly in the Rolpa and Rukum districts in Western Nepal. Promising to fight against widespread poverty and the discrimination of minorities as well as to fundamentally transform the government system of the country, the Maoists particularly mobilized members of Nepal's marginalized communities into their ranks. Among them were women, members of the lower Hindu castes, and minorities, such as Madhesis from the southern Terai plains (Pettigrew and Shneiderman 2004;Kantha 2011). What began as a low intensity conflict as the Maoists were initially only fighting the Nepal Police (NP), escalated into a conflict with a much higher number in annual battle-related deaths from 2001 onwards. In response to attacks on the RNA by the Maoists' armed wing, the People's Liberation Army (PLA), the king mobilized the army in 2001, and the government additionally created the paramilitary Armed Police Force (APF) to defeat the insurgency. In November 2006, the Maoists and a coalition of seven political partiesbut not the royal palacesigned a peace agreement and agreed on several political and military reforms, including to disarm, demobilize, and reintegrate the combatants of the PLA.
DDR thus represented a key component of Nepal's peace process. Within weeks following the signing of the accord, Maoist combatants moved into seven cantonment sites and 21 smaller satellite camps that were set up throughout the country. The camps were established with the help of the small monitoring UN Mission in Nepal (UNMIN). In April 2007, UNMIN reported that all weapons were registered and stored in sealed containers monitored by its personnel, although by the terms of the peace accord the keys to these containers stayed with the on-site PLA commanders and were not completely detached from the combatants (United Nations 2007). Only in 2012, the Maoists allowed the NA to take over the control of its weapons, thereby completing the disarmament process in a more narrow understanding of the concept (Dahal 2012). Demobilization and reintegration were equally protracted until 2012-2013. Among the issues that prolonged these phases were political disagreements between the Maoist leadership, the leaders of the political parties and the NA on mechanisms regarding the reintegration of child soldiers, the number of verified combatants, monetary compensation for ex-combatants, and on the mode of integrating Maoist ex-combatants into the NA (INT-02, 24.04.2017). In the end, more than 15,000 ex-combatants chose the 'retirement package' which entailed cash payments (Bogati 2015), while 1400 ex-combatants joined the ranks of the NA (Martin Chautari 2013; Subedi 2015). The cantonments closed in 2013.
Today, most observers regard Nepal's DDR process as one of the key accomplishments of the peace process and stress that the Maoists have successfully transformed from an armed group into a political party that has wonand lostelections in the postconflict period (Ishiyama and Batta 2011). 1 Can this be explained by the configuration of veto players? We defined veto players above as those actors that possess the skills and ability to veto the implementation of a DDR program. Between 2007 and 2013, the period of implementing DDR in Nepal, the number of veto players was select. Because the royal palace dramatically lost its influence in the initial days of the transition in 2006-07, the number of veto players is effectively limited to (1) the Maoist rebels that later (and after the end of the DDR program) split into several factions (cf. below), (2) the government side to the conflict that itself consists of several political parties and coalitions, as well as (3) the NA.

The number of veto players
This select number of veto players helped the successful implementation of the DDR process in Nepal. The only rebel group that was to be subjected to DDR was the CPN (M). At the time of signing the CPA, this Maoist rebellion was approximately 30,000 fighters strong, although this number included about 3000 child soldiers and 1000 combatants that had been recruited after the signing of a ceasefire accord in mid-2006(United Nations 2007. While the PLA chiefly fought with small arms, it had demonstrated that it could effectively control the countryside and bring significant parts of territory under its control (International Crisis Group 2005).
The government, particularly the Nepali Congress (NC) party and (to a lesser extent) the United Marxist-Leninist (UML), controlled the NP and APF security agencies during the war and thus in total approximately 100,000 officers. The NAthat had increased from 46,000 to 96,000 soldiers by the end of the warhad traditionally been under de facto control of the royal palace (Adhikari 2015). But even though it was brought under parliamentary control in 2006 and is an institution that abides by the decisions of the civilian leadership, it is regarded as a major player in the post-conflict period with significant power to disrupt the peace process (Sotomayor 2014). 2 The palace itself, however, was sidelined in the early days of the peace process and could no longer represent a major threat to peace or the DDR process, hence losing its position of a veto player. On the one hand, this loss of power for the palace eased cooperation between other players (INT-17, 30.09.2015, INT-18, 06.10.2015. This is also highlighted by an ex-deputy commander of the PLA, who points out that all parties 'were in a common agreement and […] on the people's side, [while …] the monarchy [… was] the representative of some minor […] feudal class. So if there had been a […] coup, it wouldn't have been successful' (INT-01, 26.09.2015). One the other hand, the sidelining of the palace and its inability to take on the role of a veto player was also the result of cooperation between the Maoists and the political parties. The former knew that they could not capture the state on their own, for instance as the political parties had significant support in the cities (Gobyn 2009). And the political parties -fighting the increasingly authoritarian tendencies of the Kingrealized the benefits of bringing the Maoists into their collective protests against the royal palace (Subedi and Bhattarai 2017). In sum, the small number of veto players in Nepal thus provides evidence for our first hypothesis, as it aided the cooperation between parties and the successful implementation of the DDR process.

The distance between veto players
Another important factor in the successful implementation of Nepal's DDR process was the small distance between players. This becomes evident by two means. First, even though there existed an ethnic component to Nepal's waras the Maoists chiefly mobilized members of marginalized ethnic and social groups into their ranks (Gellner 2007)parties did not split along ethnic lines. Instead, all leaders of the Maoists, government parties, and security agencies belonged to the identity group that has always dominated the political system (von Einsiedel, Malone, and Pradhan 2012). All were male, Nepalispeaking, high-caste Hindus from Nepal's central hill region. Ethnic or religious minorities, lower castes, women, or individuals from the northern Himalayas or the southern Terai region of Nepal were only represented in lower ranks of the parties or security agencies (Adhikari 2015). Many leaders of the different warring parties had therefore established close relationships before the war, also while studying at universities in India (Jha 2014). In interviews, a former government minister and NC member thus pointed out that lines of communication always remained open (INT-06, 19.10.2015).
While this dominance of one identity group has contributed to a resurgence in minority politics and protests in recent years (Sijapati 2013), it aided the implementation of DDR because it translated into a small distance between veto players. This is because the belonging to similar identity groups eased communication, rebuilding of trust, and cooperation on an elite level, and especially between PLA commanders and NA generals in the Special Committee for the Integration and Rehabilitation of Maoist Combatants, one of the main bodies entrusted with the DDR process. Mediators involved in the formal and informal negotiations regarding the DDR process and the Special Committee remembered that after initial hesitations, the warring party representatives started to trust each other more and more (INT-14, 03.05.2017). And former NA generals tended to point out they felt 'more similar than different' to their Maoist counterparts; and that by working together in the Special Committee they became 'brothers' (INT-07, 27.09.2015, INT-08, 25.04.2017. Second, what also helped the successful implementation of DDR was that the parties' policy agendas with regard to the reintegration of ex-combatants were not opposed, again showing the small distance between players. At the leadership level, and due to the high degree of local ownership in the peace process, the parties could not point to the international community for any delays in implementing the DDR process. Instead, they were fully aware that voters would blame any failures in implementation on the parties themselves (Suhrke 2011). A civil society expert also noted that Maoist leaders also wanted to ensure their control over state institutions not only through the electoral process, but also by integrating as many of combatants as possible into the NA (INT-02, 24.04.2017). Despite initial resistancearguing that the inclusion of politically indoctrinated combatants into the NA would pose a risk to the army's professionalismthis goal was supported by NA leaders (INT-08, 25.04.2017). In interviews, an academic and a security expert reasoned that the NA wanted to demonstrate its commitment to peace to the UN, for which it traditionally serves as one of the largest troop contributors (INT-05, 02.05.2017, INT-14, 03.05.2017. Prestige and reputation are further reasons cited by NA generals, who argue that as members of UN peacekeeping forces they had themselves overseen DDR programs abroad and had felt 'humiliated' if they could not have completed this process at home (INT-07, 27.09.2015).
At the rank-and-file level, also Maoist combatants and NA soldiers were keen in bringing the DDR process to an end. Having spent their youth in war, ex-combatants hoped for a speedy realization of the DDR process as many of them wanted to get married and start families (INT-09, 10.10.2015, Martin Chautari 2013. Many ex-combatants also settled in urban areas as opposed to their often rural places of origin, as urban areas (and especially Kathmandu) offered improved access to jobs, education, and infrastructure (Bogati 2015).
NA soldiers perceived that only a successful DDR process meant they could return to visit their families in villages that had been controlled by the PLA in war (FG-01, 28.04.2017). Evidence from Nepal's DDR process thus provides support for our second hypothesis: the small distance between players due to (1) a lack of ethnicized relationships and (2) a presence of similar policy interests contributed to the successful implementation of the DDR process.

The fractionalization of veto players
While the Maoist movement can be characterized as coherent and with a strong leadership under Pushpa Kamal Dahal (Prachanda) and Baburam Bhattarai at the start of the DDR process, it became increasingly horizontally and vertically fractionalized over time.
On the one hand, there is evidence that this increasing fractionalization even positively impacted DDR implementation, contrary to our theoretical expectations. Horizontally, the unified Maoist movement led by war-time leader Prachanda experienced several splits in the aftermath of the war. A first splinter group was formed towards the end of the DDR process in 2012 by Mohan Baidya (Kiran) and Ram Bahadur Thapa (Badal); other factions led by Netra Bikram Chand (Biplav) and Baburam Bhattarai broke away in 2014 and 2015 respectively. Vertically, the Maoist movement became also less cohesive in the post-conflict period, as the gap between leadership and combatants widened. Several observers account this disintegration to the level of corruption among elites. They remark that as soon as leaders joined the transitional government, they underwent a 'remarkable change in … lifestyle' (Dahal 2008, 28) by using their new role in state institutions to prioritize their individual economic benefits over the needs of combatants (Adhikari and Gautam 2014). One political opponent and member of the NC noted that Maoist elites became 'the richest people' of 'the richest party within one year' (INT-10, 12.10.2015). This alienated elites and combatants, as the latter felt betrayed by the leadership (INT-11, 22.09.2015, INT-12, 05.10.2015. In a widely circulated interview, Prachanda's ex-driver was quoted: 'The ideals that we have fought for have all been wasted … This is not the communist spirit. This is why I decided to disassociate from Prachanda' (Nepali Times 2012).
There is some evidence that this disintegration aided the implementation of the DDR process, and a former child soldier noted that the ex-combatant's disillusionment with the leadership prevented any hypothetical return to arms (INT-03, 02.05.2017). One former NA general held: Many people ask me if [ex-combatants] will take up their weapons again. My answer is [they] will not … because they were so badly used by their political leaders. They gave their lives, but they got nothing in return. Will they again go to war to make somebody rich, make somebody Prime Minister? No. (INT-13, 29.09.2015) On the other hand, it is difficult to establish a causal link, as the dynamics of the DDR process may have caused this increasing fractionalization in the first place. This becomes evident by looking at concrete measures taking place in the demobilization camps. For instance, an aid worker remembered setting up radios in the camps to provide combatants waiting for their demobilization with leisure activities (INT-15, 12.11.2015). But, as a civil society leader pointed out, only after these radios were permitted, combatants became exposed to other political opinions than those of their own leaders, began to follow the political negotiations in Kathmandu, and started to become frustrated by the leadership's behaviour (INT-16, 19.10.2015). Evidence from Nepal thus does not provide support for our third hypothesis: a decreasing cohesion of veto players did not affect the successful implementation of DDR.

Veto players and post-conflict DDR in the DRC
The DRC (former Zaire) suffered from extensive exploitation and destruction during colonial times, and in 1965 almost seamlessly transitioned into the authoritarian regime of one of the most corrupt dictators on the African continent, Mobutu Sese Seko. Mobutu's Zaire was a safe haven for many rebel groups, such as the Ugandan Lord's Resistance Army (LRA) and the Allied Democratic Forces (ADF). Due to this undisturbed activity of countless rebel groups, insecurity grew in the whole Great Lakes region (cf. Prunier 2009). In 1994, the genocide in Rwanda heralded the fall of Mobutu. The influx of hundreds of thousands of refugees from Rwanda and growing insecurity led to the development and intervention of the Alliance des Forces Démocratiques pour la Libération du Congo-Zaire (AFDL), supported by Rwandan, Ugandan, and Angolan troops who aimed to terminate the insecurity originating from Zaire. After a seven-month long war, the alliance overthrew Mobutu in May 1997 and Laurent-Desiré Kabila took power in his place. Kabila soon fell out with his former allies from Rwanda, Uganda, and Burundi, and they supported rebel groups that fought against the new government, such as the RCD and MLC (cf. below). On Kabila's side, his allies from SADC, Angola, Namibia and Zimbabwe, intervened in the violent conflict. Sudan and Chad also intervened briefly on the side of Kabila, but they withdrew in 1999. Because of the involvement of nine different countries and countless rebel groups the war between 1998 and 2002 is also known as 'Africa's World War' (Prunier 2009).
A withdrawal of foreign troops was agreed upon in the Lusaka agreement of 1999, yet it was not until 2002 that all foreign regular forces left the country. At the same time, many rebel groups from Uganda, Burundi, or Rwanda were (and partly are) still active in the DRC, competing against the countless Congolese rebel groups that were the source of insecurity particularly in the east of the country. The main aim of the DDR program was thus to reduce the number of armed groups in the country and to end the violence.
Shortly after the Lusaka peace agreement in 1999, the United Nations Organization Stabilization Mission in the Democratic Republic of the Congo (MONUC, renamed MONUSCO in 2010) was established with the aim to support peace-building measures in the country. At the official end of Congo's Second war in 2002, after foreign troops had officially left the country, the World Bank started the biggest and most comprehensive DDR program in its history, the Multi-Country Demobilization and Reintegration Program (MDRP) (cf. World Bank 2010). It ran from 2002 to 2009 and included five national programs and ten special projects. In the DRC, the lead program was the Programme National de Désarmement, Démobilisation et Réinsertion, PNDDR. At that time, the DRC had virtually no capacity in place for handling the programme, and so the Commission Nationale de la Démobilisation et Réinsertion (CONADER) was formed as a body to manage the DDR program (Scanteam 2010, 75f.). By the end of 2008, a total of 102,014 of the Congolese ex-combatants were officially demobilized (against a target of 150,000) (World Bank 2010, 28f.), and demobilization continued throughout 2009. All of the 102,014 ex-combatants received reinsertion payments, and cell phone payments were used to compensate for the lack of a functioning banking system. Actual assistance for reintegration was provided to 54% of the demobilized combatants (ibid.). Those fighters who have not been demobilized continue to be active in the DRC, for example in the countless Mai-Mai factions across the country.
In general, since 2002 peace and stability have slightly improved in the DRC. However, it is doubtful that this can be attributed to the PNDDR, but rather to a change in power dynamics in the country. The demobilization of the over 100,000 ex-combatants can be attributed to a successful political processes that led to the signing and subsequent implementation of peace agreements (Scanteam 2010, 119). However, after the end of the DDR program, violence was still ongoing particularly in the eastern parts of the DRC, and countless armed groups as well as Congolese regular forces are still responsible for insecurity and instability in the country.
During the DDR process, the DRC faced a high number of veto players from inside and outside the country. The main actor in the process was the Congolese government. After the killing of his father Laurent-Desiré Kabila in 2001, Joseph Kabila took over as president. During the peace process, he tried to co-opt some of the veto players into the government by giving them important positions and control over ministries. Particularly the Rassemblement Congolais pour la Démocratie (RCD) and the Mouvement pour la Liberation du Congo (MLC), the two major rebel groups fighting against the government between 1998 and 2003, were co-opted into Kabila's government. The different fractions of RCD summed up to about 70,000 fighters, whereas the MLC counted about 30,000 fighters (Allen 2011, 48). The armed elements of RCD and MLC were incorporated into the Congolese regular armed forces, the Forces Armées de la République Démocratique du Congo (FARDC), and the elites received influential posts in the government.
Another major veto player are the Congolese regular armed forces, or FARDC, summing up about 100,000-200,000 soldiers by the official end of the war (Allen 2011, 48). While they were in theory controlled by the Congolese government, they became in fact more and more heterogonous during the peace process due to several waves of integration of different rebel groups. Particularly in the periphery of the country, they tend to have a life of their own, and are responsible for serious human rights violations (cf. HRW 2010).
The Front des Nationalistes et Intégrationnistes (FNI) was a rebel group that was mainly active in Ituri and was dominated by ethnic Lendus. There are no figures about the number of fighters of the FNI. It is said that they were supported by Rwanda and had close ties to AngloGold Ashanti, a corporate mining company that cooperated with the FNI on gold extraction (Allen 2011, 41). The FNI participated in several rounds of DDR, and the arrest of FNI leadership in February of 2008 lead to a dissolution of the movement. Another Ituri-based rebel group and opponent of FNI was the Union des Patriotes Congolais (UPC). The UPC drew its fighters from the ethnic Hemas and it is said that the UPC was supported by Rwanda (Allen 2011, 42). The majority of its roughly 15,000 fighters has been processed through the DDR program.
The Mai-Mai militias continue to be a major veto player across the DRC until today. A lose network of combatants fighting against the infiltration of Congo by 'foreign forces', they underwent significant DDR measures. However, because they are so widespread and splintered, the militias continue to be active as of today. In 2004, the Mai-Mai militias were estimated at 30,000-50,000 fighters (World Bank 2009, 2).
Besides the more genuine Congolese armed groups there were also countless foreign groups active on Congo's territory. One of the main veto players in the East of the DRC during the 2000s were the Forces démocratiques de libération du Rwanda (FDLR) with an estimated strength of 6000-8000 fighters (Allen 2011, 43). They are the remnants of the Hutu militias that were responsible for the genocide in Rwanda in 1994 and that had fled the country after the Rwandan Patriotic Front intervened to protect the Tutsis. As they are not seen as Congolese armed group, they are not integrated into the national DDR program. As a result of the ongoing violence against Tutsis and Rwanda by the FDLR, Laurent Nkunda, a former FARDC commander, established the Tutsi-led Congrès national pour la défense du peuple (CNDP). According to estimates by Congolese military authorities in Goma, the CNDP consisted of about 8000-8500 fighters in 2007, although these estimates are disputed (HRW 2007, 21). It is said that the CNDP was largely backed by the Rwandan government (HRW 2007, 4), but was active mostly in the East of the Congo. The activities of CNDP were brought to an end with the arrest of Laurent Nkunda by Rwandan forces in January of 2009. CNDP forces were reincorporated into the FARDC.
During the 2000s, two Ugandan rebel groups were active in the Northeast of Congo: the Allied Democratic Forces (ADF) and the Lord's Resistance Army (LRA). There is no reliable information on the number of fighters for both groups (cf. Allen 2011, 51;Pehle and Speyer 2015). Both groups continue to be a source of insecurity and violence. The Burundian Conseil National Pour la Défense de la Démocratie-Forces pour la Défense de la Démocratie (CNDD-FDD) used to be active in the war from 1998-2003. With the signing of the peace agreement in Burundi in 2005 it became a major political party and retreated from Congo. Another Burundi rebel group fighting alongside the Congolese military against the Burundi army was the Forces nationales de libération (FNL). During the peace process in Burundi, they also laid down their arms and demobilized in Burundi.
Besides these 'obvious' veto players there are other actors who were indirectly linked to the actors in Congo. For instance, even after they withdraw from the DRC, Uganda and Rwanda continued to support anti-Kabila rebel groups with weapons and funding.

The number of veto players
As becomes evident from the description in the previous section, the sheer number of veto players already posed a challenge to the successful implementation of the DDR program. While some of the Congolese armed groups such as RDC and MLC could be co-opted into the government or were in such a weak position to participate voluntarily in the DDR program, it was particularly those groups that had a regional link that prevented a successful implementation of the programme. For instance, the FDLR was not integrated into any of the five national programs of the MDRP: the conflict with the government of Rwanda was ongoing throughout the 2000s. The Rwandan government refused to negotiate with those who committed the genocide in 1994 (Des Forges 1999). And from Congolese side they were seen as Rwandan rebels that should not undergo a Congolese DDR program (Autesserre 2010).
Another important veto player particularly after 2006 was the CNDP. Even though most of the fighters originated from Congo, they refused to demobilize. The CNDP saw their aim in fighting against the Rwandan rebels from the FDLR, and they continued to receive military support from the Rwandan government, which made them more likely to be veto players, according to veto player theory (Cunningham 2013, 38). Other foreign rebel groups were only slowly demobilized (such as the Burundian CNDD-FDD and FNL). The Ugandan LRA and ADF continue to be active in Northern Congo and the Central African Republic as of 2017 (Uppsala Conflict Data Program 2017).
The number of veto players from within and outside of Congo provides empirical evidence for our first hypothesis: the high number of rebel groups who did not participate in the DDR program decreased the likelihood of successfully implementing the programme.

The distance between veto players
Another factor that might have influenced the implementation of the DDR process in Congo was the distance between the different veto players. While the different splinter groups, for instance of RDC and MLC, did not have very diverse policy preferences, it was particularly ethnic conflicts that lied at the heart of the polarization between the Congolese warring parties.
This becomes evident particularly in the Tutsi-Hutu divide that has a destabilizing impact on the whole region until today. Again, the Hutu militias of the FDLR were not seen as genuine Congolese, and thus they did not participate in the Congolese DDR program. However, they continued to be a source of insecurity not only for Rwanda, but for Eastern Congo as well. As a result, the largely Tutsi-led CNDP was founded and fought against the FDLR in the East of the Congo. Rwanda continued to support the CNDP, as they saw the FDLR as major threat to their own security, but also to members of the Rwandophone community in Congo (Autesserre 2006, 7). However, for Rwanda the continued activities of the FDLR were not the only reason to support anti-FDLR forces, but also a good excuse to engage in economic activities in the resource-rich East of Congo and the exploitation of valuable resources as well (Global Witness 2005;Autesserre 2006, 8). Autesserre argues that 'the combination of these security, political, and economic interests led Rwanda to actively support several armed groups responsible for local violence in the eastern Congo' (Autesserre 2006, 8). And also national actors such as the RCD or Mai-Mai took ethnic tensions as an excuse for continued warfare (Autesserre 2006, 11).
Hence, particularly in the Congolese setting the continued violence cannot be alone reduced to ethnic rivalries and the (ethnic) distance between veto players. While rebel groups usually recruit their combatants from specific ethnic background, the conflicts are often caused by competition over land and resources. As Autesserre states, 'local antagonisms over land and traditional power led to violence' particularly in the east of the country (Autesserre 2006, 3). This includes competition over land and resources, poverty, and a lack of a governing or controlling authority (Autesserre 2006, 13ff.;Stearns, Verweijen, and Baaz 2013, 30). Militia networks drove and were driven by the development of a war economy, which thrived on illegal taxation, smuggling, and racketeering (Stearns, Verweijen, and Baaz 2013, 21). While this economy allowed for the quick enrichment of some, millions of civilians depended on it for survival, leaving them little choice but to collaborate with armed groups. Further, over the course of the peace process, the increasing availability of many ex-combatants did exacerbate this situation: they were often not re-integrated into their communities or did not have any alternative means of gaining a livelihood (Stearns, Verweijen, and Baaz 2013, 77). A re-mobilization was hence easier or the only feasible option for the former fighters to survive.
Evidence from the Congolese DDR process does not fully support our second hypotheses that the likelihood of successfully implementing post-conflict DDR programs decreases as the policy distance between veto players increases.

The fractionalization of veto players
The last factor that played a decisive role in the implementation of the DDR program in Congo was the fractionalization of the Congolese veto players. The government, for instance, was due to the integration of different conflict parties highly fractionalized. Several actors in the government such as RCD and MLC opposed DDR of their military wings particularly in the first years until the election. The different parties wanted to keep their command and control structure until elections had been held (Scanteam 2010, 66f.). However, even after the elections and once the government was in place, the motivation towards DDR continued to be reluctant. As mentioned before, control over land and resources particularly in the periphery of the country was too lucrative for some armed groups to give up arms and participate in the DDR process. Further, as Stearns, Verweijen, and Baaz (2013, 8) argue, 'the implied logic of this process-granting insurgents political power in order to quell their insurgencies-persists until today, creating incentives for elites to mobilize armed groups'. Thus, the co-optation of former warring parties into the government or military provided further incentives for mobilization instead of ending the violence.
Further, the FARDC, Congolese regular armed forces, were also highly fractionalized, which impeded the peace process and lead to a renewed outbreak of violence. By repeatedly integrating armed groups into the FARDC, the government has not only provided incentives for further insurrection, it has effectively sanctioned impunity (Stearns, Verweijen, and Baaz 2013, 9). As stated before, some of the brigades led a life of their own without clear command from Kinshasa, and committed serious human rights violations particularly in the east of the country (cf. HRW 2010). After the elections in 2006 the brigades of Laurent Nkunda broke away from the Congolese forces with the aim to protect the Congolese Tutsis from the violence of FDLR. The newly built CNDP did not only fight against the FDLR, but also against other forces of FARDC, and committed serious crimes against non-Tutsi civilians (HRW 2007).
Non-state armed actors were also highly fractionalized, which contributed to a failure of the DDR program. One example are the Mai-Mai militias, which are a lose and non-cohesive network of armed groups fighting against foreign groups in Congo (HIIK 2017, 70f.). They are spread across different regions of the Congo. The fractionalization of the Mai-Mai militias exacerbates their demobilization; due to their widespread nature, more localized demobilization and reintegration programmes would be necessary to address the different claims by different factions.
Evidence from the Congolese DDR process thus supports our third hypotheses that the likelihood of successfully implementing post-conflict DDR programs decreases as the cohesion within collective veto players decreases.
In Table 1, we give an overview of the findings from both case studies.

Conclusion
Under what conditions are DDR programs successfully implemented after intrastate conflict? Evidence from Nepal and the DRC shows that veto players can play a major role in fostering or preventing a successful implementation of DDR. However, results are ambiguous concerning different aspects of the veto player theory. First, our findings show that the number of veto players is a sound explanation for the (lack of) implementation of DDR in both cases. In Nepal, the small number of veto players helped the successful implementation of DDR, as it was beneficial to the Maoist rebels' integration into the post-conflict political system and favored cooperation between players in the DDR program. On the contrary, the high number of veto players from within and outside the country in Congo posed a particular challenge to the implementation of the DDR program. While donors tried to tackle the regional dimension of the conflict with a Multi-Country Demobilization and Reintegration Program, it was impossible to integrate all players and demobilize the high number of diverse groups. Thus, our first hypothesis can be confirmed: The likelihood of successfully implementing post-conflict DDR programs decreases as the number of veto players increases.
The findings in regard to our second hypothesis on the distance between veto players are not that straightforward. In Nepal, similar policy goals and a lack of ethnic adversary were key to the successful implementation of the DDR program, as it benefited cooperation and trust-building. In Congo, the Tutsi-Hutu divide might be seen as the heart of the ongoing violence in the country. However, this is only part of the story: the root causes of conflict in Congo lie in the competition over land and resources, and are exacerbated by poverty and the lack of a governing authority. Hence, the ethnic component of the conflict is not sufficient to explain the failure of DDR, and we do not find sufficient evidence for our second hypothesis that the likelihood of successfully implementing DDR decreases as the policy distance between players increases.
Third, evidence from Nepal indicates that a growing fractionalization of the rebel group towards the end of the DDR program even aided its successful implementation, as the alienation of rank-and-file combatants and rebel leaders prevented a remobilization for violence. In Congo, on the other contrary, the strong fractionalization of all important playersthe government, the military, and the rebel groupsimpeded a successful implementation of the DDR process. Low: The parties did not split along ethnic lines and had similar policy interests (e.g. on army integration).
High: The Hutu-Tutsi divide has a destabilizing impact; however, continued violence is also caused by competition over land and resources. Cohesion High (at first): The CPN (M) was a unified rebel group in 2007, but its fractionalization increased during (and due to) the DDR process.
Low: All main actors (government, military, and rebel groups) were highly fractionalized.
DDR program Successfully completed 2007-2013: The CPN (M) has not rearmed and remobilized but has transformed into a political party that has won and lost elections.
Ongoing violence: Between 2002 and 2009, the MDRP demobilized over 100,000 combatants, but the root causes of conflict were not addressed.
These findings are highly relevant as they link scholarship on post-conflict DDR and peacebuilding with the theoretical assumptions of the veto player theory. We adapted the veto player approach to explore and reveal mechanisms and aspects that affect the implementation of DDR programs. The two case studies offer rich empirical insights into key components of the theoretical approach, including the number of veto players, their distance, as well as their internal cohesion. The case studies provided a hard test for the main assumptions of the theoretical approach by applying it to a challenging environment of a highly dynamic and volatile post-conflict countries (George and Bennett 2005, 23f.), and clarified some of the main mechanisms that are specific to the study of implementation of DDR programs in two post-conflict countries. We are thus able to gain novel and detailed insights on determinants of effective DDR programs and post-conflict peacebuilding.
We also revealed two limitations of the veto player approach. First, Tsebelis developed the theory to analyze political behaviour in stable political systems, but particularly the DRC shows that conditions are different in highly volatile post-conflict environments. Even as players agreed to participate in DDR, the success of implementation was not always given. Due to the political and socioeconomic context, former combatants were often unable or unwilling to stick to peace, and took up arms again. This shows that a deal to disarm and demobilize is not sufficient if other political or socioeconomic challenges are not tackled simultaneously. Second, veto player theory also shows weaknesses when it comes to dynamic post-conflict environments and a change in the constitution of veto players. It may be assumed that in Nepal, a high fractionalization at the start of the peace processes would indeed have prevented a successful implementation of DDR, as too many fractions would perhaps have prevented a joint agreement. However, as we demonstrated above, the growing fractionalization of the rebel group towards the end of the DDR program in fact aided the implementation. Such diverging effects of the independent variable throughout time are hard to model and cannot be assessed as straightforward as assumed by the theory.
These findings leave room for future research, for example on the impact and embeddedness of veto players in a particular political or socioeconomic that affect preferences of players and policy outcomes alike. In addition, future models can improve the theoretical conceptualization of behaviour and preferences of veto players in highly dynamic and volatile political situations that are so often found in conflict and post-conflict environments.